Investor Relations and Retail Market Reaction: Text Mining and Event Study of Online Stock Forums in Taiwan

Author : ChihYung Wang, Tsai-Ching Liu, Peng-Wei Yang and Min-Chi Chuang

Abstract : Corporate conference calls are a primary mechanism for Investor Relations Management (IRM) aimed at reducing information asymmetry. However, in retail- dominated emerging markets, the effectiveness of such disclosures is increasingly filtered through social media sentiment. This study examines the interplay between online investor sentiment and stock performance surrounding conference calls of listed electronics companies in Taiwan. Utilizing text mining techniques on the PTT Stock Board—the region’s most influential financial forum—we construct a sentiment index to analyze investor reactions over a three- day window postevent. By employing an event study methodology, we assess the impact of positive, negative, and neutral sentiment on Cumulative Abnormal Returns (CAR). Our empirical results reveal significant market inefficiencies and sentiment asymmetry. First, we observe significant pre-event price reactions for the full sample and specifically within the positive and neutral sentiment groups, suggesting potential information leakage or strong market anticipation prior to the official disclosure. Second, post-event trajectories diverge significantly: while positive sentiment groups sustain an upward trend, negative sentiment groups— which show no pre-event leakage—experience immediate and sustained negative abnormal returns. These findings highlight the critical role of online sentiment in interpreting IR activities and suggest that in retail-heavy markets, social media discussions serve as a rapid, albeit occasionally preemptive, channel for information incorporation.

Keywords : Investor Relations Management (IRM), Sentiment Analysis, Conference Calls.

Conference Name : International Conference on Global Banking Systems and Innovation (ICGBSI - 26)

Conference Place : Osaka, Japan

Conference Date : 3rd Feb 2026